Kang Suhyun commited on
Commit
4c0bb84
1 Parent(s): 9f68c4f

[#25|#128] Add Llama 3.2 models and available models list (#136)

Browse files

* [#25|#128] Add Llama 3.2 models and available models list

This PR adds Llama 3.2 models using DeepInfra API.

Also, it adds a list of available models to the UI to help users check what models are available.

* 002

Files changed (2) hide show
  1. app.py +3 -0
  2. model.py +6 -5
app.py CHANGED
@@ -103,6 +103,9 @@ with gr.Blocks(title="Yanolja Arena", css=css) as app:
103
  - After you see both results, pick which one you think is better.
104
  """)
105
 
 
 
 
106
  with gr.Row():
107
  category_radio = gr.Radio(
108
  choices=[category.value for category in response.Category],
 
103
  - After you see both results, pick which one you think is better.
104
  """)
105
 
106
+ with gr.Accordion("Available Models", open=False):
107
+ gr.Markdown("\n".join([f"- {model.name}" for model in supported_models]))
108
+
109
  with gr.Row():
110
  category_radio = gr.Radio(
111
  choices=[category.value for category in response.Category],
model.py CHANGED
@@ -146,17 +146,18 @@ supported_models: List[Model] = [
146
  Model("gpt-4o-2024-08-06"),
147
  Model("gpt-4o-mini-2024-07-18"),
148
  AnthropicModel("claude-3-5-sonnet-20240620"),
149
- VertexModel("gemini-1.5-pro-001",
150
  vertex_credentials=os.getenv("VERTEX_CREDENTIALS")),
151
- VertexModel("gemini-1.5-flash-preview-0514",
152
  vertex_credentials=os.getenv("VERTEX_CREDENTIALS")),
 
 
153
  Model("meta-llama/Meta-Llama-3.1-8B-Instruct", provider="deepinfra"),
154
  Model("meta-llama/Meta-Llama-3.1-70B-Instruct", provider="deepinfra"),
155
  Model("meta-llama/Meta-Llama-3.1-405B-Instruct", provider="deepinfra"),
 
 
156
  Model("Qwen/Qwen2.5-72B-Instruct", provider="deepinfra"),
157
- Model("Qwen/Qwen2-72B-Instruct", provider="deepinfra"),
158
- Model("google/gemma-2-9b-it", provider="deepinfra"),
159
- Model("google/gemma-2-27b-it", provider="deepinfra"),
160
  ]
161
 
162
 
 
146
  Model("gpt-4o-2024-08-06"),
147
  Model("gpt-4o-mini-2024-07-18"),
148
  AnthropicModel("claude-3-5-sonnet-20240620"),
149
+ VertexModel("gemini-1.5-pro-002",
150
  vertex_credentials=os.getenv("VERTEX_CREDENTIALS")),
151
+ VertexModel("gemini-1.5-flash-002",
152
  vertex_credentials=os.getenv("VERTEX_CREDENTIALS")),
153
+ Model("google/gemma-2-9b-it", provider="deepinfra"),
154
+ Model("google/gemma-2-27b-it", provider="deepinfra"),
155
  Model("meta-llama/Meta-Llama-3.1-8B-Instruct", provider="deepinfra"),
156
  Model("meta-llama/Meta-Llama-3.1-70B-Instruct", provider="deepinfra"),
157
  Model("meta-llama/Meta-Llama-3.1-405B-Instruct", provider="deepinfra"),
158
+ Model("meta-llama/Llama-3.2-3B-Instruct", provider="deepinfra"),
159
+ Model("meta-llama/Llama-3.2-1B-Instruct", provider="deepinfra"),
160
  Model("Qwen/Qwen2.5-72B-Instruct", provider="deepinfra"),
 
 
 
161
  ]
162
 
163